review_3 of heart disease prediction using ecg images.pptx

heart disease prediction using ecg images

1. Comparative Analysis of Deep Learning Architectures for Cardiovascular Disease Detection from ECG Images Review-1 Department Of ECE BY(Batch.No-221) Bodige Nikhila-160121735072 Tulluri Kinnera-160121735088 Hrithik Kenche-160121735103 Supervisor- Dr.Naga Devi Assistant Professor Dept Of ECE 1

2. Motivation of the Work Our team chose this area due to the significant global burden of cardiovascular diseases (CVDs) and the critical need for early and accurate detection. ECGs are a widely used tool for diagnosing heart conditions, and we saw an opportunity to apply deep learning to enhance this process. By exploring different architectures like SqueezeNet, AlexNet, CNN and XceptionNet.We aim to compare their efficiency in identifying various heart abnormalities. This project combines our collective interest in AI with its potential to make a real impact in healthcare, improving diagnostic precision and, ultimately, patient outcomes. 2

3. Area of the Work DEEP LEARNING TECHNIQUES CONVOLUTIONAL NEURAL NETWORKS SQUEEZE NET ALEX NET XCEPTION NET MACHINE LEARNING ALGORITHMS  Support Vector Machine KNN Random Forest Decision Tree Naive Bayes 3

4. PROBLEM STATEMENT Cardiovascular diseases are a major global health concern. We leverage pretrained deep learning models like SqueezeNet and AlexNet , along with Machine Learning models to predict cardiac anomalies. Traditional machine learning methods complement our approach. It excels in achieving better early diagnosis of cardiovascular diseases. 4

5. OBJECTIVES Utilize a dataset of electrocardiogram (ECG) images of cardiac patients for model training and evaluation, emphasizing its importance in real-world applicability. Develop a predictive model for cardiovascular diseases using deep learning and Machine Learning techniques.  Evaluate the effectiveness of pretrained model networks as feature extractors for conventional machine learning algorithms. Explore the use of traditional machine learning algorithms, including support vector machine, K-nearest neighbors, decision tree, random forest, in conjunction with pretrained models as feature extraction tools. 5

6. SYSTEM ARCHITECTURE 6

7. MODULE DESCRIPTION Module 1: Data Preparation and Preprocessing Objectives: Prepare the ECG image dataset for feature extraction and classification. Ensure data consistency and quality. Methods: Conversion to grayscale. Image resizing. Grid line removal. Benefits: Consistent data format for analysis. Noise and irrelevant information removal. Proper data division for model training and validation. 7

8. Module 2: Feature Extraction and Data Engineering Using AlexNet and SqueezeNet Objectives: Extract meaningful features from ECG images. Convert images into structured data for analysis. Dataset split for training and testing. Approach: Utilize deep learning models (AlexNet and SqueezeNet). Benefits: Improved model performance through meaningful feature extraction. Simplified, one-dimensional data representation. 8

9. Module 3: Classification Using Machine Learning Models (SVM, KNN, Random Forest) Objectives: Automatically classify ECG images into relevant classes. Evaluate the performance of different machine learning models. Methods: Employ Machine Learning models (SVM, KNN, Random Forest,Naive bayes,Decision Tree). Train models on the feature-engineered dataset. Classify ECG images into four classes. Evaluate model performance using metrics. Select the best-performing model. Benefits: Automation of ECG image classification. Ability to choose the best-performing model. 9

10. Software And Tools Used VSCode ANACONDA Software Deep Learning Models CNN Squeeze Net Alex Net XceptionNet Machine Learning Algorithms 10

11. Datasets The dataset contains 4 different classes of ECGs: (Normal, Myocardial infarction, Abnormal Heart beat, History of Myocardial infarction). Total number of ECGs in each category are: (Normal - 284, Myocardial infarction - 239, Abnormal Heart beat - 233, History of Myocardial infarction - 172) Link: https://data.mendeley.com/datasets/gwbz3fsgp8/2 11

12. Results: ( Squeezenet ) 12

13. CNN 13

14. Alexnet 14

15. Xception 15

16. Summary 16

17. References [1] M. Swathy and K. Saruladha , “A comparative study of classification and prediction of cardio-vascular diseases (CVD) using machine learning and deep learning techniques,” ICT Exp., to be published, 2021. [Online]. Available: https://doi.org/10.1016/j.icte.2021.08.021 [2] S. Kiranyaz , T. Ince, and M. Gabbouj , “Real-time patient-specific ECG classification by 1-D convolutional neural networks,” IEEE Trans. Biomed. Eng., vol. 63, no. 3, pp. 664–675, Mar. 2016. [Online]. Available: https://doi.org/10.1109/TBME.2015.2468589 [3] Shahil Sharma, Rajnesh Lal, and Bimal A Kumar. "Machine Learning for Early Detection of Cardiovascular Disease in Fiji." 2023 IEEE Asia-Pacific Conference on Computer Science and Data Engineering (CSDE) , DOI: 10.1109/CSDE59766.2023.10487655. [4] Emine Yaman , Ali Almisreb , Mehmet Akif Tanisik , and Nooritawati Md Tahir. "Diagnosis of Cardiovascular Diseases Using Classification Algorithms." 2022 IEEE Symposium on Wireless Technology & Applications (ISWTA) , DOI: 10.1109/ISWTA55313.2022.9942793. [5] Meirzhan Baikuvekov , Azhar Tursynova , and Galymzhan Yespayev . "A Deep Learning for Cardiovascular Diseases Detection on Wearable Devices Data." 2024 IEEE 4th International Conference on Smart Information Systems and Technologies (SIST) , DOI: 10.1109/SIST61555.2024.10629525. [6] Damodar Prabhu K , Prathiksha Rao “ Detection and Analysis of Cardiovascular Diseases using Machine Learning Techniques ,” Scientific reports, vol. 11, no. 1, p. 8886, 2021. 17

18. [7] M. Naz, J. H. Shah, M. A. Khan, M. Sharif, M. Raza, and R. Damaševiˇcius , “From ECG signals to images: A transformation based ap- proach for deep learning,” PeerJ Comput . Sci., vol. 7, 2021, Art. no. e386, doi : 10.7717/peerj-cs.386. [8] A. H. Khan, M. Hussain, and M. K. Malik, “Cardiac disorder classification by electrocardiogram sensing using deep neural network,” Complexity, vol. 2021, 2021, Art. no. 5512243. [Online]. Available: https://doi.org/10.1155/2021/5512243 [9] A. H. Gonsalves, F. Thabtah , R. M. A. Mohammad, and G. Singh, “Pre-diction of coronary heart disease using machine learning: An experimental analysis,” in Proc. 3rd Int. Conf. Deep Learn. Technol., 2019, pp. 51–56. [Online]. Available: https://doi.org/10.1145/3342999.3343015 [10]M. B. Abubaker and B. Babayiğit , "Detection of Cardiovascular Diseases in ECG Images Using Machine Learning and Deep Learning Methods," in IEEE Transactions on Artificial Intelligence, vol. 4, no. 2, pp. 373-382,April2023,doi:10.1109/TAI.2022.3159505. 18

19. Thank you 19